Question

我有两个2个熊猫数据帧：

timestamp1 = ['2018-10-01 00:01:49.800000000', '2018-10-01 00:01:52.900000000', '2018-10-01 00:04:18.857741600'] 

df1 = pd.DataFrame(timestamp1, columns =['timestamp']) 

timestamp2 = [['2018-10-01 00:01:50.230 ', 'John'], ['2018-10-01 00:01:52.560', 'Jill'], ['2018-10-01 00:04:19.100', 'Jack']]

df2 = pd.DataFrame(timestamp2, columns =['timestamp', 'name'])

我想基于时间戳（t）合并两个帧，其中t来自df1（t）> = df2（t）。我正在寻找的输出是：

         timestamp_df1,               timestamp_df2,                name
2018-10-01 00:01:49.800000000      2018-10-01 00:01:50.230          John
2018-10-01 00:01:52.900000000      2018-10-01 00:01:52.56           Jill
2018-10-01 00:04:18.857741600      2018-10-01 00:04:19.100          Jack

时间戳是我可以合并的数据帧中唯一常见的东西。我已经研究了条件合并，但似乎这不是正确的路线？任何帮助或建议都将是有用的！

Answer 1

您似乎需要在这里pd.merge_asof。请注意，尽管在第二种情况下条件不成立。您也可以像在这里一样添加公差：

pd.merge_asof(df1, df2, 
              on='timestamp', 
              direction='nearest', 
              tolerance=pd.Timedelta('1min'))

     timestamp  name
0 2018-10-01 00:01:49.800000000  John
1 2018-10-01 00:01:52.900000000  Jill
2 2018-10-01 00:04:18.857741600  Jack

时间戳合并数据帧大于或等于

1 个答案: